Parallel governance institutions built by Maoist rebels during Nepal's civil war led to lasting improvements in political participation among historically excluded groups, more representative candidate selection across parties, and greater local government capacity.
Since 1945, more than 200 revolutions have challenged existing regimes, and over half have rewritten the formal rules of politics (Beissinger 2022). Yet their long-run legacies differ dramatically: some collapse into renewed authoritarianism or state failure, while others produce durable, inclusive democracies (Albertus and Menaldo 2014, Matanock 2017). In recent work (Bhusal et al. 2026), we examine whether one part of the revolutionary process explains these divergent trajectories: the parallel governments rebels build while fighting the state.
We study this question in Nepal, where the Maoist People’s War (1996–2006) transformed a 240-year caste-based monarchy into a federal democratic republic. During the war, the Maoists established sub-national ‘People’s Governments’ across large parts of the country. We show that these People's Governments durably mobilised historically excluded groups into politics, shifted post-war candidate selection in local elections to be more representative, and improved local government capacity.
A revolution for federal democracy and against entrenched exclusion
Nepal is one of the most ethnolinguistically diverse countries in the world (Drazanova 2019). In addition, social exclusion was formally codified before the People’s War. The 1854 civil code enshrined a Hindu caste hierarchy in which upper-caste Bahun, Chhetri, and Newar groups (BCNs; 45% of the population) dominated political and economic life, while Indigenous Janajatis (38%) and Dalits (13%) were systematically marginalised. In 1995, on the eve of the war, BCNs had a poverty rate of 25%, compared to roughly 50% for Janajatis and Dalits.
This entrenched exclusion motivated the Maoists, then a banned political party, to launch the People’s War in 1996. Their 40-point demands, submitted on the eve of the war and rejected by the government, called for the end of monarchy and the creation of a secular federal democracy (Hutt 2004). The war began in Maoist strongholds and expanded outward; by 2006, roughly half the country was under some form of Maoist control. The 2006 Comprehensive Peace Accord ended the conflict, laid the foundations for the federal democratic constitution adopted in 2015, and allowed the Maoists to enter formal democratic politics.
In addition to demanding a new political order after the war, Maoists had begun constructing elements of one in areas they controlled during the war. These wartime institutions were the People’s Governments: Maoist-led sub-national administrations formed in territories under rebel control. They replaced the central state in these areas, running courts, collecting taxes, and challenging exclusionary practices. They were also meant to advance prototypes for the federal democracy that followed. As part of this programme, they cultivated leaders from historically excluded groups: the Maoists barred elites from contesting wartime elections and placed Janajati cadres in positions of authority (Sharma 2004, Lawoti and Pahari 2010).
Isolating the effect of exposure to rebel governance
Maoist expansion was not random. They moved, for example, into places that were rural, forested, and favourable to guerrilla warfare. Hence, simply comparing places with and without People’s Governments would not identify the effects of rebel governance. To address this, we exploit the fact that People’s Governments were organised around historical district boundaries drawn decades before the war and often following natural barriers. We compare citizens and municipalities located close to district borders, where one side fell under People’s Government rule, and the other did not. These nearby areas looked similar before the war in forest cover, education, ethnic composition, population density, and other characteristics, making this a credible comparison for studying the impacts on political mobilisation and local state capacity.
Political mobilisation and increased engagement of Janajati citizens
Using a nationally representative geocoded survey from Nepal’s first federal elections in 2017, we show that People’s Governments increased post-war political knowledge and participation, but only among the historically excluded Janajatis.
Janajatis in People’s Government districts are 29 percentage points more likely to know about the 2015 Constitution, 32 percentage points more likely to know about party activities, 25 percentage points more likely to trust political parties, and 12 percentage points more likely to participate in political campaigns. We do not find corresponding effects among upper-caste BCNs. This suggests that wartime exposure closed – and in some cases reversed – the long-standing knowledge and engagement gap between Janajatis and upper-caste citizens.
Changes in party gatekeeping dynamics
In Nepal's party-centred system, candidate-selection committees decide who gets to stand for mayor. Evaluating the process of candidate selection is difficult because party nomination processes are usually opaque. We therefore collected internal party administrative data and ran novel surveys with party leaders across 11 central districts. We assembled rosters for the three major parties, matched aspirants to census records, and administered an Implicit Association Test (IAT) – a standard psychological tool used to measure bias (Greenwald et al. 2003, Beaman et al. 2009, Carlana 2019) – to selection committee members.
In this empirical analysis, we compare parties and ethnic groups within the same local political environments to ask whether deeper exposure to People’s Governments changed who parties selected, while holding fixed many features of the electorate. These novel data show that moving from partial to full People’s Government districts increases the probability that a selection-committee member is Janajati by 26 percentage points. Maoist Janajati committee members additionally show less implicit bias against Janajati leadership than their BCN counterparts in their committees, a pattern absent in non-Maoist committees.
More inclusive political selection across parties
We also see a striking effect of People’s Government exposure on non-Maoist parties that were not organising the revolution. In districts where the Maoists were less present, non-Maoist parties nominate Janajatis at barely half the rate of BCNs. In districts where the Maoists established consolidated People’s Government, the gap nearly disappears. Candidate nomination closely represents the population in these districts for all parties.
Electoral outcomes follow. In partial People’s Government districts, BCNs are twice as likely as Janajatis to win mayoral office. In full People’s Government districts, Janajatis are twice as likely to win as BCNs. The share of Janajati mayors also rises from 35% to 54%. These patterns sharpen in the 2022 election, suggesting an equilibrium shift towards more inclusive politics.
More state capacity in exposed local governments
Inclusion is not the only legacy. Two decades after the war, municipalities in People’s Government districts score higher on a wide range of administrative-capacity measures covering procurement procedures, infrastructure, service delivery, disaster management, judicial functions, and fiscal capacity. People’s Government municipalities also receive 13% more in conditional federal grants.
One channel is who governs. Using an event-study design around the 2022 local elections, we compare municipalities where Maoist mayors replaced non-Maoist mayors to those without such a change. Administrative capacity rose sharply after Maoist mayors took office, and these gains were concentrated in districts exposed to Maoist governance during the war. This is consistent with continuity of leadership from wartime governance into the post-war state.
Wartime institutions can reconfigure political equilibria
Nepal’s People’s Governments were dissolved after the war. Yet the institutions they built – and the leaders, networks, and civic capacities they created – left a measurable imprint nearly two decades later. In formerly exposed areas, parties became more inclusive in whom they nominated, Janajatis became more likely to hold office, and local governments scored higher on administrative capacity while attracting more federal grants, with no evidence of weaker fiscal discipline.
This pattern helps explain why Nepal differs from cases where rebel governance left citizens alienated from the post-war state. In El Salvador, for example, recent evidence finds that rebel governance generated lasting distrust, disengagement, and worse long-run economic outcomes (Bandiera et al. 2022). Nepal’s post-war settlement took a different form: the peace process dismantled the unitary monarchy, opened the way to inclusive federalism, and brought the Maoists into competitive democratic politics on terms that reflected their demands.
Revolutions impose enormous human and economic costs, and their legacies range from repression and renewed conflict to democratic consolidation. A central question in development economics is how inclusive and capable institutions arise. Nepal illustrates one pathway: rebel institutions mobilised excluded groups and cultivated new leaders, while the post-war settlement channelled those legacies into competitive democratic politics rather than leaving them outside the state. The result was not only more descriptive representation. Janajatis came to hold office roughly in proportion to their population share, and Janajati politicians were as positively selected on observable characteristics as upper-caste politicians. In this case, inclusion and state capacity advanced together, suggesting that sustained efforts to broaden political representation need not come at the expense of effective governance.
References
Albertus, M, and V Menaldo (2014), "Gaming democracy: Elite dominance during transition and the prospects for redistribution," British Journal of Political Science, 44(3): 575–603.
Bandiera, A, L I Dinarte Diaz, J M Jimenez, S V Rozo, and M M Sviatschi (2022), "Rebel governance and development: The persistent effects of guerrillas in El Salvador," NBER Working Paper.
Beaman, L, R Chattopadhyay, E Duflo, R Pande, and P Topalova (2009), "Powerful women: Does exposure reduce bias?" Quarterly Journal of Economics, 124(4): 1497–1540.
Beissinger, M R (2022), The Revolutionary City: Urbanization and the Global Transformation of Rebellion, Princeton University Press.
Bhusal, B, M Callen, R Pande, S A Prillaman, D Singhania, and A Subedi (2026), "From people's war to people's rule: Rebel governance and the foundations of inclusive democracy," NBER Working Paper.
Carlana, M (2019), "Implicit stereotypes: Evidence from teachers' gender bias," Quarterly Journal of Economics, 134(3): 1163–1224.
Drazanova, L (2019), "Historical Index of Ethnic Fractionalization Dataset (HIEF)."
Greenwald, A G, B A Nosek, and M R Banaji (2003), "Understanding and using the Implicit Association Test: I. An improved scoring algorithm," Journal of Personality and Social Psychology, 85(2): 197.
Hutt, M (2004), Himalayan People's War: Nepal's Maoist Rebellion, Indiana University Press.
Lawoti, M, and A K Pahari (2010), The Maoist Insurgency in Nepal: Revolution in the Twenty-First Century, Routledge.
Matanock, A M (2017), Electing Peace: From Civil Conflict to Political Participation, Cambridge University Press.
Sharma, S (2004), "The Maoist movement: An evolutionary perspective," in M Hutt (ed.), Himalayan People's War, Indiana University Press.